does the fix nph in lammps stands for isoenthalpic-isobaric ensemble?I have done a simulation to verify this.
but the result makes me disapointed . Am I wrong? If it’s not
isoenthalpic-isobaric ensemble, How can I realize it in lammps.thanks in advance.
Material Science And Engineering Department
Xi’An Technological University
NPH in LAMMPS is constant pressure, but no temperature control,
so it's like a subset of NPT. If it's not what you want (look at the
details in fix_nph.cpp), you'd have to write your own fix_foo.cpp
that does the ensemble you desire.
2009/6/18 limitworld <[email protected]>:
Well spotted. press/berendsen does indeed slowly draw energy out of the system, hence nve + press/berendsen can not be considered as a reasonable approximation to the NPH ensemble. This is in contrast to nve+temp/berendsen and nve+temp/berendsen+press/berendsen, which provide reasonable approximations to NVT and NPT ensembles, respectively. It is important to note that the Berendsen methods are known to be ad hoc, and they do not conform to any known ensembles. The LAMMPS documentation is consistent with this view.
I have been running long NPH simulations of Beryllium with a solid-liquid interface in order to obtain an estimate of the melting point. I have found that fix nph works well (no energy drift), provided that the timestep is small enough and provided also that the drag keyword is not used. Using a drag term causes behavior similar to what you saw with press/berendsen.
2009/6/22 Steve Plimpton <[email protected]…29…>